Hi, I've a question about protein docking. In the manual for protein docking, there is another application for preparing structures for docking: docking prepack protocol. This prepack protocol can generate input structures from bound protein-protein complex. So how can I generate input structures from unbound protein components (for example, if protein A and B form a complex, and I have the structures for A alone and B alone)? Right now, I just do a fast relax for A and B, and then combine them into the same pdb file, then use docking prepack to generate the input structure. Does this make sense? Also, for some complex structures, after prepack, I got positive scores. I suppose this is normal since the interface sidechain rotamers are not optimized. Is this correct? Also, sometimes when I run docking from unbound components of the complex, I get this error message: ERROR: Conformation: fold_tree nres should match conformation nres. conformation nres: 628 fold_tree nres: 639 When I comment out the native complex structure, there is no more error. So I checked the native complex pdb and the separate component pdbs and found some residues are missing in either the native complex pdb or the component pdbs. Is there a way to add these missing residues back to the pdb so I can use the native complex pdb to make calculation directly? Thanks for help.